Participants
A sample of 127 (65 Spanish and 62 Colombian) individuals with ASD and their families participated in a survey. Participants came from all regions of Spain and Colombia. Participating children were enrolled at centres delivering special education, early childcare or day care, or regular schools with open classrooms. Spanish participants came from different-sized urban and rural areas in the regions of Valencia, Murcia and Andalusia. The Colombian sample came from the Andean, Caribbean, Orinoquía and Pacific regions.
DSM-5 diagnostic criteria for ASD [1] served as the referent for inclusion of individuals in both the Spanish and Colombian samples. Spanish participants had previously been diagnosed by mental health services and pertinent institutions responsible for ascertaining the degree of disability and dependency. Similarly, autism diagnoses in Colombian participants were made by public and private mental health institutions (e.g., early intervention centres, autism centres, etc.).
Individuals with other concomitant diagnoses, such as motor disabilities, multiple disabilities, attention-deficit/hyperactivity disorders, obsessive-compulsive disorders, neurodegenerative diseases and mental illnesses, were excluded. Understandably, participants with ASD and ID, were included when ASD was the primary diagnosis.
Measures
Sociodemographic questionnaire
Lam and Aman's [16] sociodemographic questionnaire was adapted for the present online study. This tool consists of a series of questions about sociodemographic characteristics (e.g., age, sex, country of birth) and comorbidities (e.g., intellectual disability).
Clinical questionnaire on gastro-intestinal symptoms
This is an ad hoc questionnaire that was developed to identify gastro-intestinal disorders according to Rome criteria [2]. It consists of a series of questions about gastrointestinal disorders (e.g. diarrhoea, abdominal pain, dyspepsia and gastroesophageal reflux).
Gastrointestinal Symptom Severity Scale (GSSS)
This instrument is based on Rome IV criteria [2] and consists of seven items pertaining to main gastro-intestinal symptoms (constipation, diarrhoea, average stool consistency, stool odour, flatulence and gas, and abdominal pain). The instrument comprises an abdominal subscale (abdominal pain, gas and constipation) and a vomiting and defecation subscale (vomiting, defecation in inappropriate places, diarrhoea, rumination). Items are rated along a four-point Likert scale ranging from 0 (none/nothing or this symptom does not occur) to 3 (very frequent and troublesome). The GSSS presents adequate psychometric properties in individuals with autism and in neurotypical children and adolescents [17, 18, 19]. Internal consistency coefficients of .73 have been reported in children with typical development [19], whilst coefficients between .75 and .61 have been reported in individuals with autism [18]. Two versions of the instrument are available, namely, a version for caregivers-professionals and a self-report version. The caregivers-professionals version of the test was administered in the present study.
Pain and sensitivity reactivity scale (PSRS)
The PSRS is a tool that evaluates reactivity to pain and sensory reactivity according to 50 items. It is composed of three dimensions, namely, pain, sensory hyporeactivity and sensory hyper-reactivity. Items are rated on a four-point Likert scale ranging from 0 (behaviour does not occur) to 3 (behaviour occurs and is a severe problem). Both hyposensitivity and hypersensitivity dimensions comprise tactile, olfactory, visual, gustatory and auditory items. In addition, the PSRS includes a pain reactivity domain which comprises seven items. The PSRS was elaborated based on theoretical requisites conceived by Miller et al. [20], which characterizes sensory modulation disorders according to three patterns (hyper-response, hypo-response and sensory seeking) as proposed nosology for diagnosis. Two versions of the PSRS are available. The first is a version administered for completion by caregivers and professionals, whilst the second is a self-report version [21]. The caregiver version of the PSRS has shown excellent internal consistency in samples with ASD (pain α = .83; broad sensory hyporeactivity α = .90; broad sensory hyperreactivity α = .93) [22]. The caregiver version was used in the present study.
Procedures
The present study was approved by the Ethics Committee of the University of Alicante in Spain (reference: UA-2019-10-04) and the Catholic University of Pereira in Colombia (reference: UCP-2021-05). The survey was carried out using online data collected from the parents and caregivers of individuals with ASD who provided written informed consent. The reporting assessment protocol was individually applied through online survey tools, specifically, LimeSurvey (LimeSurvey GmbH, Hamburg, Germany), in Spain, and Formsite, in Colombia. Both platforms complied with the same comprehensive autism assessment protocol.
For the recruitment of all participants, a cover letter containing study information was sent to ordinary centres, special education centres, associations serving families with children with autism, etc. Centre management teams informed families about the study and provided access to the online survey. Centres held an online or telematic meeting in order to timetable appointments and explain the purpose of the research. Subsequently, participating institutions contacted families to organize a meeting and further explain the purpose of the study. Similarly, some institutions provided the research team with contact information so that study researchers could directly explain the purpose of the study to families. Finally, an explanatory video was shared using social networks. All participating families and caregivers had a child diagnosed with ASD according to DSM-5 criteria [1]. Individuals with ASD, whether with or without ID, were diagnosed according to DSM-5 criteria using standardised scales (e.g., Wechsler nonverbal scale of ability, Leiter-3 scale, etc.). Participants had all been previously diagnosed by the pertinent mental health services and institutions responsible, in their given country, for establishing their degree of disability and dependency. Families with children with another type of neurodevelopmental disorder (e.g., ADHD) or with a diagnosis of ASD but who had not been assessed for possible ID were excluded from the study.
With regards to individuals with ASD in Spain, diagnoses were made at early care centres and through regional paediatric services attached to local mental health centres [23]. Common standardised scales were used to reach a diagnosis of autism (e.g., communication domain of the Vineland adaptive behavior scales, Vineland adaptive behavior scales, autism diagnostic interview-revised, autism diagnostic observation schedule–generic, childhood autism rating scale, etc.). However, the diagnostic process for individuals with ASD in Colombia was applied in line with the "clinical protocol for the diagnosis, treatment and comprehensive care route for boys and girls with autism spectrum disorders" outlined by the Colombian Ministry of Health [24]. This national protocol uses applied behaviour analysis (ABA) to reach clinical and non-psychometric diagnoses. A limitation of this protocol is that most of the standardized scales used to apply it have not yet been validated and scales taken adapted from measures used with Spanish or American populations are typically used. Study researchers organised a training session for all participating centres, in which they describe in detail the purpose of the research study, the tests to be used and the process to be followed during test administration.
Expert psychologists and similar professionals (educational psychologists, special education teachers, psychologists) were engaged in the administration of online tests to participating families, which sought to gather information on the perceptions and knowledge of individuals with ASD at the participating institutions that served them. Psychologists based at participating centre were on hand to help families resolve any doubts regarding their diagnosis in the first part of the survey. The explanatory video mentioned above urged families to consult pertinent psychological and psychiatric reports should any doubts exist regarding their diagnosis.
The European Spanish versions of the GSSS and PSRS were revised by three specialist Colombian psychologists and one Spanish psychologist who corroborated to ensure that items were culturally equivalent. Subsequently, clarity and ease of understanding of all items were verified in a pilot study conducted with ten participants in Pereira (Colombia). This revealed that no issues emerged regarding understanding and so no linguistic adaptation was needed.
Data analyses
Statistical analyses were performed using IBM SPSS Statistics v 25.0 for Windows [25]. Means and standard deviations for all items corresponding to all tests were calculated for both samples from the direct scores reported by participants. Cronbach's alpha scores were calculated for all individual subscales and overall scores pertaining to the GSSS and PSRS, as a function of country. Next, as a preliminary step prior to performing the main analysis, sociodemographic variables were compared between the two countries in order to examine potential confounding factors. Chi-square analyses were conducted of the categorical variables of gender, age and comorbidity with intellectual disability (ID). For the comparison of mean values, the Kolmogorov-Smirnov test was applied to verify the normality assumption for the use of parametric tests within each group. Outcomes revealed that the normality assumption was not satisfied. As a result, the non-parametric Mann-Whitney U test was used to determine whether GSSS and PSRS subscale and overall scores differed between Spanish and Colombian participants, with significance being set at p < .05. Effect sizes associated with these differences were calculated to determine whether statistically significant differences existed between the proportions reported in each country. In this sense, .20 ≤ d ≤ .50 represented a small effect size, whilst .51 ≤ d ≤ .79 corresponded to a medium effect size and d ≥ .80 reflected a large effect size [26].